API Changes for February 24, 2009

[Matt Sanford]

We launched a few different things today, 2009-02-24:

* Fixed: Calls to /friendships/exists.json were returning
incorrect results. (http://code.google.com/p/twitter-api/issues/detail?id=312
)
* Feature: The /users/show call now accepts the parameters
user_id and screen_name to allow disambiguation of numeric screen
names. (http://code.google.com/p/twitter-api/issues/detail?id=162)

And for those in the OAuth closed beta:

* Security (OAuth): Misconfigured applications could use insecure
PLAINTEXT signatures. This has now been explicitly disallowed.
* Fixed (OAuth): Changed OAuth tokens to work correctly during
database replication delays.
* Feature (OAuth): Removed the approval process for new
applications. New applications will be allowed by default and we will
review and revoke as needed.

Thanks;
— Matt Sanford

[Dimebrain]

I think there's some erroneous caching occurring with these optional
parameters for disambiguation when using JSON.

If I make this request:
http://twitter.com/users/show.json?screen_name=413

I get back the correct user whose screen name is 413 and whose id is
16089382.

If I then make this request according to the API wiki:

http://twitter.com/users/show.json?user_id=12345

I receive my previously requested user with screen name 413 and id
16089382.

This only happens when sending JSON requests, it appears to work as
expected with XML.

[Matt Sanford]

Hi there,

    I see this for both XML and JSON. Looking into it now. If I wait a few moment (10 seconds or so) it works correctly, which points me toward the correct cache. I'll let you know when I find something.

Thanks;

  — Matt Sanford

[Matt Sanford]

Hi there,

    I found the source of the issue and have a fix ready. The issues is a 1 minute cache elsewhere in the system. I'm awaiting a code review now and hope to make it in time for today's deploy to fix it.

Thanks;

  — Matt Sanford

On Feb 25, 2009, at 05:39 AM, Dimebrain wrote:

[Matt Sanford]

Hi there,

    It seems GMail has been eating some of my outgoing mail today so you may not have seen my last reply. This has been fixed.

— Matt

On Feb 25, 2009, at 05:39 AM, Dimebrain wrote: